RL-Native Distillation: Exploiting Scored Trajectories for Few-Step Image Generation

• URL: http://arxiv.org/abs/2608.09226v1
• Date: Mon, 10 Aug 2026 07:49:05 GMT
• Title: RL-Native Distillation: Exploiting Scored Trajectories for Few-Step Image Generation
• Abstract summary: Efficient text-to-image generation requires both reinforcement-learning (RL)-based reward alignment and few-step distillation.<n>We propose REST (Reward-Enhanced Scored-Trajectory Distillation), a single-stage RL-distillation co-training framework.<n>We show that REST enables few-step CFG-free inference that matches or surpasses its 40-step RL teacher.
• Score: 38.215412661414604
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Efficient text-to-image generation requires both reinforcement-learning (RL)-based reward alignment and few-step distillation, yet these procedures are typically performed sequentially, increasing training cost and risking the loss of reward gains during compression. We instead take an RL-native perspective: diffusion RL already generates reward-scored finite-step trajectories, whose intermediate states provide a natural source of distillation supervision rather than a disposable byproduct of sampling. Based on this insight, we propose REST (Reward-Enhanced Scored-Trajectory Distillation), a single-stage RL-distillation co-training framework that attaches a decoupled student to an arbitrary RL teacher. The student learns segment-wise from the teacher's evolving rollout trajectories while leaving the original teacher optimization unchanged. To prevent uniform imitation from preserving undesirable low-reward behaviors, we further introduce Advantage-Modulated Distillation (AMD), which transforms rollout advantages into signed weights over a base distillation loss. AMD strengthens supervision from preferred trajectories and mildly repels the student from low-reward ones. The resulting framework is lightweight and plug-and-play, requires no extra image rollouts, no separate distillation dataset, and no adversarial training. Experiments on compositional generation, visual text rendering, and human-preference alignment show that REST enables few-step CFG-free inference that matches or surpasses its 40-step RL teacher, with an overall additional training cost below 25% over pure RL. REST improves DrawBench PickScore over RTDMD by 0.82 while requiring only one-fifth of the training iterations.

Related papers

• Distill Where You Fail: Recovering Learning Signals of Negative RL-Groups from Adaptive Teacher Guidance [50.77757355236912]
  We propose RSTG (Recovering Learning Signals via Adaptive Teacher Guidance), which applies distillation selectively and precisely where it matters most.<n> Experiments demonstrate that RSTG substantially outperforms naive GRPO+OPD by +4.02% on math and+3.05% on code.
  arXiv  Detail & Related papers  (2026-08-01T17:29:14Z)
• Distilled Reinforcement Learning for LLM Post-training [8.435213882597546]
  Large language model (LLM) post-training is essential for improving reasoning, adaptation, and alignment.<n>Existing methods mainly follow two paradigms: reinforcement learning (RL) and on-policy distillation (OPD)<n>We propose Distilled Reinforcement Learning (Distilled RL), which integrates teacher supervision into the RL objective to provide fine-grained guidance.
  arXiv  Detail & Related papers  (2026-07-19T13:32:54Z)
• RationalRewards: Reasoning Rewards Scale Visual Generation Both Training and Test Time [51.256131853751754]
  We show that teaching reward models to produce explicit, multi-dimensional critiques transforms them from passive evaluators into active optimization tools.<n>We introduce Preference-Anchored Rationalization (PARROT), a principled framework that recovers high-quality rationales from readily available preference data.<n>The resulting model, RationalRewards (8B), achieves state-of-the-art preference prediction among open-source reward models.
  arXiv  Detail & Related papers  (2026-04-13T15:38:09Z)
• Diffusion Reinforcement Learning via Centered Reward Distillation [35.979608265594685]
  We present textbf Reward Distillation (CRD), a diffusion RL framework derived from KL-regularized reward models built on forward-process fine-tuning.<n>To enable reliable text-to-image fine-tuning, we introduce techniques that explicitly control distribution drift.<n>Experiments on text-to-image post-training with textttGenEval and textttOCR rewards show that competitive SOTA reward optimization results with fast convergence and reduced reward hacking preference.
  arXiv  Detail & Related papers  (2026-03-14T21:29:33Z)
• Reinforcement-aware Knowledge Distillation for LLM Reasoning [63.53679456364683]
  Reinforcement learning (RL) post-training has recently driven gains in long chain-of-thought reasoning large language models (LLMs)<n>Most existing knowledge distillation methods are designed for supervised fine-tuning (SFT), relying on fixed teacher traces or teacher-student Kullback-Leibler (KL) divergence-based regularization.<n>We propose RL-aware distillation (RLAD), which performs selective imitation during RL -- guiding the student toward the teacher only when it improves the current policy update.
  arXiv  Detail & Related papers  (2026-02-26T00:20:39Z)
• Flash-DMD: Towards High-Fidelity Few-Step Image Generation with Efficient Distillation and Joint Reinforcement Learning [32.32567390728913]
  Diffusion Models have emerged as a leading class of generative models.<n>Timestep distillation is a promising technique to accelerate generation, but it often requires extensive training and leads to image quality degradation.<n>We introduce Flash-DMD, a novel framework that enables fast convergence with distillation and joint RL-based refinement.
  arXiv  Detail & Related papers  (2025-11-25T17:47:11Z)
• Reinforcement Learning Teachers of Test Time Scaling [21.551446057221185]
  Key use case of reasoning LMs is to act as teachers for distilling new students and cold-starting future RL iterations.<n>We introduce a new framework that avoids RL's exploration challenge by training a new class of Reinforcement-Learned Teachers (RLTs)<n>RLTs are prompted with both the question and solution to each problem, and tasked to simply "connect-the-dots" with detailed explanations tailored for their students.
  arXiv  Detail & Related papers  (2025-06-10T02:53:24Z)
• Towards Better Alignment: Training Diffusion Models with Reinforcement Learning Against Sparse Rewards [52.90573877727541]
  reinforcement learning (RL) has been considered for diffusion model fine-tuning.<n>RL's effectiveness is limited by the challenge of sparse reward.<n>$textB2text-DiffuRL$ is compatible with existing optimization algorithms.
  arXiv  Detail & Related papers  (2025-03-14T09:45:19Z)
• Improved Distribution Matching Distillation for Fast Image Synthesis [54.72356560597428]
  We introduce DMD2, a set of techniques that lift this limitation and improve DMD training. First, we eliminate the regression loss and the need for expensive dataset construction. Second, we integrate a GAN loss into the distillation procedure, discriminating between generated samples and real images.
  arXiv  Detail & Related papers  (2024-05-23T17:59:49Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.